DINH ANH KHOI NGUYEN

Implemented pipeline automation for collecting GoTo data using Power platform and JavaScript. Developed Ames AI assistant backend using LangChain for internal document queries with RBAC within the Ames tenant. Engineered chain of thoughts, created prompt templates, and optimized hybrid search with HNSW index and reciprocal ranking using Postgres on-prem server. Deployed Ames AI assistant frontend using React built with Vite and hosted on Azure App Service. Built ingestion pipeline for various document types to generate embeddings and utilized LLM agents to ensure LLM-friendly content.

Conducted research and built an OCR API backend that outperformed Google Vision by 7.3% on Character Error Rate (CER). Developed data crawling tools and regression tests for AI services. Collaborated with sales teams and helped win recurrent contracts from major companies such as Toyota, AT&T, and Office Depot.

Developed and implemented a machine learning platform. Wrote reports, user guides, and test cases for QA and management teams. Presented new features and negotiated work items with the Vice President and Product Owners. Responsible for training various machine learning models and developing backend integrations with multiple ML cloud service providers including Azure, GCP, and AWS.